SciELO - Scientific Electronic Library Online

vol.3 special edition author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Kriterion: Revista de Filosofia

Print version ISSN 0100-512X

Kriterion vol.3 Belo Horizonte  2007


The legitimating fact in the transcendental deduction of the categories
On Dieter Henrich's reading of Kant


Dieter Henrich, leitor de Kant: sobre o fato legitimador na dedução transcendental das categorias



Christian Klotz; Soraya Nour

Translated by Christian Klotz and Soraya Nour
Translation from Kriterion, Belo Horizonte, v.48, n.115, p. 145-165, 2007.




This article reconstructs the principal moments of Dieter Henrich's work on Immanuel Kant's theoretical philosophy. Henrich seeks to clarify and regain the fundaments of Kant's theory of knowledge – from which his followers, according to him, have distanced themselves – based on the analysis of the "transcendental deduction of the categories". Firstly, Henrich investigates the proof structure of deduction, comparing the first and the second edition of Critique of Pure Reason. Secondly, he investigates, in the Kantian argument, the relationship between the identity principle of self-consciousness and objectivity. Finally, extending the comparison to Critique of Practical Reason, Henrich elucidates the program and methodology in deduction, showing that the idea of a legitimating fact, borrowed from the juridical notion of a deduction, becomes the fundamental element. We analyse the problems raised by the conception of a philosophical argument based on fundamental "facts".

Key-words: Immanuel Kant, Dieter Henrich, transcendental deduction, self- consciousness, objectivity, fact.


Este artigo reconstrói os momentos principais dos trabalhos de Dieter Henrich sobre a filosofia teórica de Immanuel Kant. Henrich procura esclarecer e recuperar os fundamentos da teoria do conhecimento de Kant, dos quais seus seguidores teriam se distanciado, a partir da análise da dedução transcendental das categorias. De início, Henrich investiga a estrutura da prova na dedução, comparando a primeira e a segunda edição da Crítica da Razão Pura. Em seguida, Henrich investiga no argumento kantiano a relação entre o princípio de identidade da consciência de si, por um lado, e objetividade, por outro. Por fim, estendendo a comparação à Crítica da Razão Prática, Henrich elucida o programa e a metodologia na dedução, mostrando como o "fato" legitimador se torna o elemento fundamental.

Palavras-chave: Immanuel Kant; Dieter Henrich; Dedução Transcendental; Consciência de Si; Objetividade; Fato




In 2004, Dieter Henrich published Grundlegung aus dem Ich. Untersuchungen zur Vorgeschichte des Idealismus.1 The publication, in the year of the 200th anniversary of Kant's death, of this monumental work (1740 pages) on the beginning of German idealism, stressed the Author, was a mere coincidence. This fact was, nonetheless, an occasion for Henrich to explain his reasons for undertaking this study. Kant has formulated so many questions that he inspired a philosophical production with an incomparable complexity in the following generations. However, his idealist followers, according to Henrich, even if they intended to amplify the consequences of his thought, radically transformed its fundamental principles – a process the culmination which is Hegel's philosophy. Henrich's analyses of Kant has historical and systematic grounds: it aims to recuperate the foundations of Kant's thought, which were for his first followers obscure, and which have not yet been sufficiently clarified to this day. A true "reception" of Kant is yet to occur. 

With regard to Kant's theoretical philosophy, Henrich's project focuses on the transcendental deduction of the categories, the dense argumentation of which reveals in nuce the foundations of Kant's knowledge theory. In an article from 1989 that addresses the methodology of Kantian argument, Henrich analyses the relationship between fact and legitimation, which for Kant were not absolutely separate domains.2 As with Critique of Practical Reason,3 the transcendental deduction of Critique of Pure Reason4 refers to a fact. Deduction is not defined as a chain of syllogisms, but as a juridical argument, its "proof" consists in a reference to a legitimating fact. Indeed, Henrich elucidates, while we currently call "deduction" a mere chain of syllogisms (in this sense, we tend to interpret Kant's deduction), in the 18th century, "deduction" was the name of a legal argumentative practice in which the "proof" was based on a "fact". It is according with this legal model that Kant elaborated transcendental deductions both in Critique of Pure Reason and Critique of Practical Reason.

This article reconstructs Henrich's analysis of Kant's transcendental deduction based on three of his texts, which include the results of almost two decades of his inquiries: "Die Beweisstruktur von Kants Transzendentaler Deduktion" (1973),5 Identität und Objektivität (1976),6 "Kant's Notion of Deduction and the Methodological Background of the First Critique" (1989).7 In the first, Henrich investigates the proof structure in deduction, comparing the first and second editions of Critique of Pure Reason. Here, deduction is still considered a chain of syllogisms. In the second contribution, Henrich investigates, in Kantian argument, the relationship between the identity of self-consciousness and objectivity. Finally, the third text, extending the comparison to the Critique of Practical Reason, elucidates the program and methodology in deduction, showing how the "fact" issue becomes a central question (and, at the same time, rectifying the interpretation of deduction as a chain of syllogisms).


I. The structure of the transcendental deduction of the categories

In his article on the "proof" structure of transcendental deduction (1973),8 Henrich analyses the logical relationship between the partial steps of the deduction of the categories.  Comparing the proof structure in the first and second edition of the first Critique, he shows why the argument was fully developed only in the latter.

The task of deduction is to show that the categories of understanding are a priori valid for all objects of experience, while, at the same time, limiting the theoretical use of the categories to empirical objects. In the second edition of Critique, Kant's conclusion seems to be outlined in two completely different moments. The conclusion in Section 20 does not seem to differ from the result in Section 26. In Section 20, Kant concludes that "the multiplicity in a given intuition is necessarily subject to the categories" (B143); and in Section 26, his conclusion is that "the categories (...) are (...) valid a priori for all objects of experience" (B161).

However, interpreting the conclusions of Sections 20 and 26 as two proofs of the same proposal would lead to a conflict with Section 21, in which Kant affirms that these two arguments, rather than two separate proofs, jointly constitute the proof of the deduction: "In the proposal above begins the deduction of the pure concepts (...). In what follows (§ 26)… the intention of the deduction will be completely reached" (B144/45). Sections 20 and 26, therefore, offer two arguments with significantly different results that together lead to a unified proof of the transcendental deduction, which will be called by the author a "proof in two steps". Henrich then examines two proposals based on the theory of the double-proof9: according to Adickes10/Paton11 and according to Erdmann12/de Vleeschauwer.13

Adickes and Paton mention the distinction established by Kant in the first edition of Critique between "subjective deduction" and "objective deduction". In the Preface of this edition, Kant distinguishes two aspects of transcendental deduction: the objective aspect, which makes the validity of the categories intelligible – the demonstration that the categories have validity; and the subjective aspect, which investigates the relationship of the categories with the faculties of knowledge – the demonstration of how the categories acquire validity. For Adickes and Paton, Section 20 proves the objective validity and Section 26 demonstrates the subjective conditions of application. According to Henrich, however, the distinction introduced in the first edition of Critique does not apply to the second version of the deduction: in Section 21 of the latter, Kant states that the demonstration of the objective validity of the categories will be completed in Section 26.

Erdmann and De Vleeschauwer consider two types of demonstration established in the first version of the deduction: one that begins "from the top", that is, from self-consciousness, and one that begins "from the bottom", that is, from sensible representations. The authors apply this distinction to the argument in the second edition: Section 20 would correspond to the "from-the-top" deduction and Section 26 to the "from-the-bottom" deduction. Such a proposal, observes Henrich, agrees with Section 26, but not with Section 20, because both arguments begin with the intuition of showing that the deduction can become a unitary representation only when intellectual functions are applied to it.

Henrich proposes another solution, beginning with the proposals of Sections 20, 21 and 26. In Section 20, Kant affirms that intuitions are submitted to the categories insofar as they possess unity (B143). Kant indicates this restriction (already to possess unity) with the capital letter of the indefinite article in the expression "in An intuition" (in Einer Anschauung). This was not understood by Norman Kemp Smith (p. 160 of his translation), for whom this expression means that only an intuition is submitted to the categories. However, since the indefinite article in German – "ein" – and the word "unity" – Einheit – have the same root, with the capital letter, Kant may express not the distinction of any arbitrary intuition opposed to others (singularity), but its internal unit. Thus, concludes Henrich, the result of Section 20 is valid for intuitions insofar as they contain unity. The unity of intuition can be conceived only in accordance with the categories.14

Section 21 announces that the restriction in Section 20 will be eliminated in Section 26, which shows that the categories are valid "for all the objects of our senses" (B161). According to Henrich, the function of this step is intrinsically connected with the dualism of understanding and sensibility. In fact, Henrich says,

"everything of which we can become conscious becomes accessible for some by ways  that do not depend immediately on this consciousness. According to Kant, this is the reason why the consciousness must be understood as an activity, which is always becoming conscious, and that which we call "I" because of its internal necessary unity. But this always presupposes that, first of all, something is present that must become conscious. Thus, our consciousness can be found only together with a 'passive', receptive faculty, which is distinct and, in certain aspects, opposed to the spontaneity of consciousness."15

This reference to representations given in sensibility as a faculty irreducibly different from understanding is the root of the problem in the second step of the deduction. The first step showed that all intuitions unified in consciousness are submitted to the categories as principles of their unity. However, this does not answer the question as to whether all representations given in the sensibility can be unified. According to the first step, the possibility of unifying is limited to a part of the given intuitions. The second step, argues Henrich, has the task of excluding the possibility of such disproportion between consciousness and sensibility; thus, it has the task of showing that not only the intuitions that possess unity, but all intuitions are submitted to the categories.

According to Henrich, the fact that the second step can deal essentially with a problem related to the difference between understanding and sensibility explains why Kant appeals to the results of the Transcendental Aesthetic, emphasising the notion that time and space are pure intuitions, constituting, at the same time, the formal conditions of all empirical intuitions (see B 160). Kant uses this concept in the second step of the deduction to stress that time and space are intuitions that possess unity; consequently, their unit must be conceived in accordance with the result of the first step of the deduction, that is, as a unity that is in accordance with the categories. Thus, time and space, in which all the sensible intuitions are given, are unities structured in accordance with the categories. The condition for all sensible data as such, in virtue of the unity of time and space, is to be in accordance with the unity required by understanding (see ibid.).

Henrich emphasises that the argument of the deduction essentially refers to understanding and sensibility as irreducible and cooperative faculties. This shows that the second deduction agrees completely not only with the structure of the first Critique, but also with the Kantian conception of system.16 Kant bases the transcendental philosophy on the unifying principle of the unity of self-consciousness as "the highest point" (B 134); but in a very peculiar way of arguing ("a Kantian way", says Henrich), he combines, at the same time, principles of irreducibly different faculties. There must be a unifying principle according to which understanding operates, but such a principle should not exclude other sources of knowledge; the essential difference between the two roots of our knowledge must be considered and it must become possible to establish their unity. This conception is different from empiricism – which does not have any a priori principle of unity – as well as from speculative idealism – which does not have an essential unity of originally distinct elements. Kant's transcendental deduction – but only in the second version, as Henrich emphasises – has such a conception.


II. Identity and objectivity

1) The analysis of objectivity

After having analysed the structure of deduction, in Identität und Objektivität (1976),17 Henrich analyses how the deduction of the categories intends to connect the principle of the identity of self-consciousness with objectivity.18 This requires the analysis of certain characteristics of "I think" that are not addressed in the discussion on the general notion of reflection: its Cartesian status and its relation with the judgment form. Henrich focuses on a characteristic thesis in Kant's theory of knowledge: we only know objects of experience if we use concepts guided by principles that are not derived from experience, but founded in the constitution of self-consciousness, denominated the "transcendental unity of apperception".

Two objections have been raised against this Kantian claim. The first denies that our knowledge of experience objects has non-empirical presuppositions. The second objection contests that aprioristic presuppositions of our knowledge, once admitted, may be explained by self-consciousness. However, one can find arguments against these two objections in the Kantian doctrine of objectivity, as well as in his doctrine of the self-consciousness as an identity principle.

Henrich's argumentation begins with the Kantian inquiry regarding the conditions of objectivity. According to Henrich, this already yields sufficient results to refute the empiricist theory of knowledge. Moreover, the Kantian argument that begins with self-consciousness must appeal to the analysis of the concept of objectivity. However, the discussion on Kantian analysis of objectivity shows also that this is insufficient for the intentions of the deduction of the categories; and that Kant needs to appeal to self-consciousness as the highest principle of his argument.

In accordance with the dominant epistemology of his time, Kant's inquiry regarding the conditions of objectivity presupposes that the basic data of all our knowledge on reality are "sensations" – presentations of simple qualities in a diffuse coordination in space. However, even if knowledge of objects is possible only with such data, an elementary analysis of an object concept reveals that the reference to objects involves a different dimension. The concept of object includes the condition of constancy that the sensations do not have; in particular, an object can continue to be the same, whereas the data that present it are modified. Moreover, reference to objects requires more than just the consciousness of presentations: it aims at objectivity, distinguishing true representations from "mere" representations. However, to make this distinction, the subject only can use simple data that, in and of themselves, do not present objects. In consequence, the relationships between these data – their coherence or unity – must be the constituent aspect for a representation of an object, in contrast to a "mere" representation. As sensible data by themselves cannot yield the consciousness of their unity, this aspect must appear in the activity of the knowing subject. Consequently, Kant's notion that the "synthesis" exercised by the subject constitutes the reference to objects results from an analysis of the object concept.

However, this result is still compatible with the ideal that a synthesis depends on empirically observed regularities and, consequently, follows contingent rules. The notion – essential for Kant's position – that the synthesis referring to objects demands a priori rules is justified by another aspect of objectivity: the reference to objects is exercised essentially in judgments. Thus, it can be expected that a consideration of the structural conditions of the judgment will show that objectivity requires certain a priori rules, to which the data are submitted. Therefore, reference to the judgment form is a central element in Kantian analysis of objectivity. Here, Kant's claim is that a categorical judgment involves essentially the linking of a subject concept with a predicate concept. This seems to imply that a judgment involves the "synthesis" of two aspects or properties of the object according to an a priori rule that corresponds to the relationship between the concepts in the judgement. However, as Henrich argues, the form of the judgment does not lead directly to the Kantian conception of synthesis. The reason for this is that a categorical judgment can involve a demonstrative relating to a simple quality, a "quale" (for example: "This is green"). In this case, there is no descriptive content that corresponds to the subject term; consequently, the judgment does not involve a synthesis of aspects or distinct properties of the object. All the content of the knowledge is in the predicate. For allowing such judgments, the form of the categorical judgment itself does not imply the necessity of the synthesis.

According to Henrich, however, it must also be said that judgments with demonstrative reference to qualia are very particular cases of using the form of the categorical judgment.19 In the normal use of this form, the predicate does not absorb all the object conception that the subject possesses. Rather, there is an object conception different from the predicate concept and this is expressed by the subject term of the judgment. Therefore, in the normal use of the form of the categorical judgment, the negation of a predication does not deny our entire conception of the object. Moreover, a categorical judgment typically implies that a conjunction of judgments on the same object is possible, involving different predicates of which the judgment is a member. The normal use of the form of the subject-predicate judgment is tied to logical operations in such a way that it only can fulfil its function if the characterisation of the object is different from the predicate. Thus, it is not the mere form of the categorical judgment, but the use of this form that implies the complexity of the object and, with this, the necessity of the "synthesis" for the reference to the objects. Moreover, this result involves a specific rule of the synthesis as an a priori condition of any objective knowledge: Objects must be conceived as possessing a plurality of properties and as a unity of diverse predicates. According to Henrich, this point is more clearly expressed in a reflection by Kant written probably in 1797 (N. 6.350). In this reflection, Kant explicitly connects his theory of the synthesis with the analysis of the use of the form of the categorical judgment.20

The reconstruction of the Kantian analysis of objectivity has an important result for the transcendental deduction of the categories. It shows that objectivity, because it uses the form of the categorical judgment, requires a synthesis of data in accordance with a rule, which is a necessary condition of our use of this form. However, this also shows the limits of the analysis of objectivity as a strategy for the deduction of the categories: objectivity analysis alone can establish that our knowledge demands a synthesis in accordance with a priori rules. However, this result does not clarify to what extend empirical knowledge is possible. Thus, the analysis of objectivity does not exclude the possibility that reference to objects is limited to certain occasions, that is, the possibility of a disintegrated consciousness that cannot incorporate all its experiences into the knowledge of an objective world. However, it is essential for Kant to show that the use of the categories is necessary for all conscious experiences, which, consequently, are necessarily part of the knowledge of an objective world. To show this, Henrich argues, Kant needs to found the deduction of the categories not on the analysis of objectivity, but on the consideration of self-consciousness and its identity.

2) The identity of self-consciousness

An aspect of self-consciousness is that it is simple. When I think of myself as a thinking subject of my thought, "I" in these thoughts only means that a thinker is related in the same and only way to all his or her thoughts. Thus, the consciousness of "I" itself does not yet involve the diverse properties of myself as a person. It is "simple" in the sense that it does not show any complexity or internal plurality. At the same time, the conscious subject is related to the diverse contents of his or her thought. Thus, a characteristic of subjectivity is: to be simple and, at the same time, refer to a multiplicity. The consequence is that our self-consciousness must be a unity principle ("Einheitsprinzip"): All the thoughts of which a thinker is conscious as "his or her" thoughts belong uniquely to him or her as a simple subject.

According to Henrich, however, the unity of the thoughts, as a consequence of simplicity, does not yet demonstrate a thesis which is essential to the transcendental deduction of the categories: The unity of the subject's conscious contents must be in accordance with rules that the subject knows a priori. The unit given with simplicity means, first of all, that the contents have something in common, namely, they belong to the same simple subject. This fact does not yet include relationships between these contents established by synthesis. Thus, an argument that presupposes the simplicity of the subject should show that, in order to possess the unity implied by the simplicity of "I", the contents must be unified by means of a synthetic activity functioning according to rules. For Henrich, such an argument is possible, but would only show that it must be a synthesis according to rules to which the contents must be submitted before being conscious. This would be a condition of our consciousness of the contents. We could not affirm that the rules of such a pre-conscious synthesis must be known by the subject or even be known a priori. They would only be postulated from the philosophical standpoint. However, Kant wants to show that self-consciousness involves an a priori knowledge of the synthesis rules. Therefore, Henrich argues, is necessary to base the argument of the deduction of the categories not on the simplicity of self-consciousness, but rather on another aspect, which Kant, in his criticism of the paralogisms of rational psychology, explicitly distinguishes from simplicity: the identity of "I think". 21

In the paragraphs that Henrich considers key to the interpretation of transcendental deduction, Kant affirms that the subject "could not think a priori about his or her own identity among the diversity of the representations" if he or she did not have "in front of his or her eyes" the a priori rule according to which he or she unifies all his or her representations (A 108). According to Henrich, it is the "numerical identity" of self-consciousness that Kant considers here – making it possible to establish what cannot be demonstrated from its simplicity. However, in order to understand Henrich's argument, it is necessary to investigate variants of definitions of the numerical identity of self-consciousness. It was Leibniz who introduced the "strict" concept of numerical identity. According to this concept, objects are numerically identical if they possess exactly the same properties. This concept excludes the possibility of alteration, that is, that an object can pass from one state to another in which it possesses different properties while remaining a numerically identical object. The Leibnizian conception of substance as monad, in which all its properties are given since the beginning, is a consequence of this concept. Crusius criticised this concept of numerical identity, objecting that a singular thing can remain the same throughout the change of its states insofar as its constituent properties remain unchanged. This is "the moderate" concept of numerical identity. According to this concept, the identity of an object remains in the course of different states in which it shows different properties.

According to Henrich, it is the moderate concept of numerical identity that Kant introduces in the deduction of the categories. Kant rejects the strict concept of numerical identity on which the metaphysics of Leibiniz is founded. Moreover, says Henrich, there are arguments specifically related to self-consciousness that favour the moderate concept of identity. Since the reflection, through which the consciousness of "I think" is added to a representation, must be conceived as an action of the subject that modifies his representational state, there is a difference between a state in which the subject merely represents something and the state in which the subject refers to a representation as being "his or hers". However, the alteration of his or her state by the reflection must be considered a change in the subject that numerically remains identical and, thus, must be conceived in accordance with the moderate concept of identity.22

Beyond being founded on the concept of the moderate identity of the subject, the transcendental deduction of the categories presupposes, according to Henrich, a certain notion of the epistemological status of the consciousness of this identity. In the above-mentioned passage in A 108 – the key passage for Henrich – Kant emphasises that the subject thinks of his or her identity "a priori". This Cartesian certainty of self-consciousness is an independent premise of the notion that the identity of the subject must be conceived as moderate numerical identity. According to Henrich, the a priori status is equally constitutive for the systematic question of the deduction of the categories, namely: how is it possible for a subject to be a priori conscious of his or her numerical identity in all his or her representational states? The understanding of the argument of the deduction of the categories requires the reconstruction of the best reply that can be given to this question within the Kantian theoretical framework.

3) The relationship between identity and objectivity

It is necessary to conceive that the deduction of the categories is founded on the moderate identity of self-consciousness as well as the apriority of its certainty as two independent premises. The subject has self-consciousness with Cartesian certainty and this certainty contains the consciousness of its numerical identity. Thus, everything that is implied in the thought of numerical identity and constitutes its meaning must be involved in the a priori knowledge that the subject has of himself. However, identity implies a sequence of states in the same subject. Therefore, according to the moderate concept of numerical identity, it is only in the transition from one state to another that the subject can be the same. Thus, the subject must have knowledge, prior to all experience, of what it means to pass from one state of representation to another.

According to the Henrich's reconstruction, it is at this moment that Kant introduces the concept of a rule.23 The a priori knowledge of one's own identity must include some conception of the way through which one passes from one state to another. Without this, the a priori consciousness that the identity undergoes alterations would have no content. The consciousness of the identity includes a priori knowledge of certain modes of transition. However, this knowledge must be considered in accordance with an a priori knowledge of rules under which the transitions are accomplished. A rule is the way by which actions are exercised. With this step, an intrinsic connection between self-consciousness and a priori knowledge of rules is established, and precisely this connection is basic to the formulation of the thesis that self-consciousness makes the reference to objects necessary.

The connection between the result established and the analysis of the objectivity becomes evident if we recognise that a priori rules for the transition from one state to another are, in relation to the conscious contents in these states, a priori rules of the synthesis of these states, for synthesis in accordance with a priori rules was shown to be necessary for the reference to objects in the Kantian analysis of objectivity. Thus, concludes Henrich, the Kantian argument that investigates the conditions of the consciousness that the subject has a priori to his or her numerical identity can actually establish that the reference to objects is necessary for self-consciousness. This argument depends essentially on the moderate concept of numerical identity and on the presupposition of the apriority of the consciousness that the subject has of his or her identity.

As Henrich observes, Kant does not clearly distinguish between simplicity and identity in the deduction of the categories. There is, in Kant, a non-reflected association of the two arguments. According to Henrich, this association has a reason: The deduction of the categories can be conceived as a combination of these two arguments: first, it begins with the analysis of the unity implied by the simplicity; second, when it adds the identity as condition, it shows that this unity must be conceived with regard to the conscious transitions in the identity (the subject possesses a priori consciousness of this identity). The second step deepens the conception of unity introduced by the conception of simplicity in such a way that rules of the synthesis known a priori reveal themselves to be involved in the unity of the contents implied by self-consciousness.24.


III. The juridical aspect of deduction: Legitimation through a "Fact"

1) The juridical background of the Kantian conception of "deduction"

The analysis of the deduction argument and, in particular, of the relation between the principle of the identity of self-consciousness and objective reference, are Henrich´s starting points for investigating a fundamental question: the relationship between legitimation and fact. Henrich elucidates this relationship in his article "Kant´s notion of Deduction" (1989)25, investigating Kant's methodology. Here, he observes that, as the transcendental deduction in the first Critique is not the only one, its interpretation has to be compatible with the transcendental deduction of freedom in the second Critique, which is based on a "fact of reason". Consequently, an interpretation of the transcendental deduction of the first Critique can only be successful if it is able to show that – and how – a transcendental deduction can refer to a fact (even if we intuitively contrast the concepts of validation and fact).

Therefore Henrich's reconstruction of the methodology of transcendental deduction begins with the term "deduction".26 We generally understand "deduction" as a logical procedure through which a proposition – the "conclusion" – is derived from other propositions ("premises"). However, Henrich observes, this is not the only use of the term and not even the most common in 18th century academic language. Kant uses the term is a sense explained in the first phrase of the chapter "The principles of any transcendental deduction" in the first Critique: "Jurists, speaking about rights and usurpations, distinguish in a trial the question of right (quid juris) from the question of fact (quid facti); and they ask for a proof of both. The proof of the first, which has to establish the right, or a legal demand, is what they call deduction" (A 84/B116). To think that Kant uses deduction in a logical sense and to forget the juridical meaning of the term leads to inadequate interpretations.

But why, first of all, did Kant adopt the term deduction in the juridical context, transferring it to his philosophical context? To answer this question, Henrich analyses the argumentative form of a juridical deduction, which begins with the basic division of rights into innate and acquired.27 Innate rights are inseparable from the human being as such. However, acquired rights have a specific origin in a fact. For instance, I have a title of nobility if I am the legitimate son of a certain couple; I have an academic title if I passed certain exams; and I posses something if I acquired it in a legal way. To decide whether an acquired right really exists or is only apparent, it is necessary to explain its origin, which is a fact (for instance, the fact that something was acquired by buying it). It is to this factual origin that a deduction refers. Henrich's fundamental claim is that there is a very close analogy between a juridical argumentation and the fundamental arguments of critical philosophy, which also refer to a legitimating fact. Thus, by referring to juridical deduction in the first Critique, Kant says something very important about the structure of his argument. Questions of the form "How is ... possible?" are questions about the origin of the use of certain pure concepts of understanding, an origin which must be a fact possessing validatory force.28

One might ask, however, whether this suggestion does not mix up the question of right with the question about what the facts are. The answer is that both questions aim at understanding the origin of claims, but that they do this in different ways. For instance, in the case of a inheritance, the question of right is related to a fundamental fact: the origin of the possession must be a declaration of last will which is authentic and valid. The question of fact, however, requires knowledge on how the testament was first conceived, written and preserved – that is, it refers to circumstantial facts. Even if we cannot reconstruct the entire history of the testament – the question of fact -, the question of right could still be definitively answered. Henrich claims that this also applies to the transcendental deduction of the categories. Transcendental deductions essentially refer to the origin of our knowledge. Thus, the awareness we have of our identity as thinking subjects and of the unity of space and time are the facts with which the transcendental deduction of the categories begins. It is not necessary to gain a complete understanding of the origin and the constitution of these facts, but only of those aspects that are relevant and sufficient to justify our claims to knowledge.

2) The methodology of the Deduction

The aim of the deduction is to justify the use of the categories through an investigation of the origin of their use. However, this result does not yet yield an understanding of the particular method of the deduction. How is the reference to original "facts" involved in this argument?29 To answer this question, Henrich discusses the distinction between reflection (reflexio) and investigation (examinatio), which Kant introduced in his lectures. This distinction is based on following aspects: (a) our cognitive capacities constitute a web and cannot be reduced to a single fundamental type of operation; (b) each of these capacities operates spontaneously within its proper domain; (c) in order to acquire knowledge, each operation must remain within the limits of its domain. Reflection is our mind's implicit knowledge of its operations and the principles of these operations. It accompanies all epistemic activities; thus, it is in virtue of reflection that the mind knows what kind of operation it is exercising (for instance, that it is analysing and not synthesising; calculating, and not counting). This knowledge is intrinsically involved in the exercise of the activity itself. However, it is not an exhaustive knowledge of the epistemic processes and operations, but only an awareness of the general principles upon which the activity being exercised is based.

In contrast to reflection, investigation is a deliberate activity of the mind, which makes from its operations an object of explicit consideration. Critique (and its deductions) is an investigation about claims of knowledge. It identifies the principles that orientate a discourse and their connection to fundamental facts and operations. Henrich claims that understanding the relationship between investigation and reflection is decisive to the comprehension of the methodology of transcendental deduction: Investigation is based on reflection, on the implicit knowledge of its operations and its principles in the human mind. It makes of this knowledge an explicit, albeit still partial, knowledge of these principles and the validating relations among them.

There are two important corollaries of this result: (a) Every deduction, as investigation, must proceed by arguments that refer to what the reflection reveals. It is a clarification of the awareness that one particular operation depends on another, which is more fundamental. Thus, it shows that analysis is always, on a more fundamental level, accompanied by synthetic activities, the principles of which do not originate from experience. (b) The second corollary is that Kant did not make his philosophical methodology explicit. The central question is: How can implicit knowledge be transformed into explicit knowledge? Kant introduced the concept of "preliminary judgements", which result from the tendency of reflection to conceptualise our operations and faculties. Investigation begins with such judgements – even if they turn out to be unacceptable. However, in his lectures on logic, Kant says that we have only a rudimentary comprehension of this mechanism.

Henrich's final observation on Kantian methodology reveals an important application of his doctrine on the distinction between reflection and philosophical investigation. A key concept of deduction in the first Critique is the unity of apperception, expressed by "I think". "I think", Henrich comments, is the self-consciousness that can accompany any reflection, regardless of the kind of epistemic activity to which it refers. This suggests that the principle from which the most fundamental transcendental deduction starts must possess the very same generality and unrestricted domain that characterises the process of reflection, on which the method of justification – philosophical investigation – is continuously based. Thus, there is an intrinsic relation between the key concept of the transcendental deduction in the first Critique and the methodological principle of all deductions – between reflection and investigation.


IV. The concept of 'fact' in the transcendental context: Implications and problems

Henrich's merit is to have reconstructed the Kantian argument, while, at the same time, elucidating its methodological background. According to his account, the core of the project of a transcendental deduction is the conception of principles known through reflections on which our discourse is based, possessing – in correspondence to the juridical conception – the status of validating facts. However, this conception leads us to the following question: In what sense can the conception of "legitimating facts" be transferred from its original juridical context to the quite different context of transcendental legitimation? The concept of a 'fact' cannot be exactly the same in these two contexts; all that can be said is that they are analogous. In the juridical context, 'facts' are events that are confirmed by documents and testimony. However, the principles of a transcendental deduction are qualified as 'facts' because they are constitutive of our epistemic perspective without being demonstrable as necessary for any epistemic practice as such. In the philosophical context, the factual nature of the principles means that they cannot be derived from epistemic conditions, which are more fundamental.

At this point, it becomes clear that Henrich´s reconstruction of the transcendental deduction diverges fundamentally from that given by P.F. Strawson, which combines elements of critical and analytic philosophy.30 According to Strawson, the aim of transcendental deduction is to refute Hume's sceptical empiricism. This project requires that the argument is based on premises that the sceptic would also accept. Therefore, Strawson suggests reconstructing the transcendental deduction starting from a "minimal" conception of experience that does not involve anything more than the condition that concepts are applied to the contents of given intuitions. The conditions explicitly introduced by Kant's deduction argument – in particular, the awareness "I think" – are, in Strawson's reconstruction, established as necessary implications of the minimal conception of experience. Thus, they are not ultimate "facts", but analytic implications of the concept of experience in the minimal sense, also accepted by the Humean sceptic. It is therefore not surprising that Strawson denies the factual nature of these principles in his comment on Henrich, claiming that they can by inferred from more fundamental conditions, which are implicit in the minimal concept of experience.31

Strawson's reconstruction is important because it shows how the Kantian argument can be conceived as a refutation – a reduction ad absurdum – of Humean empiricism. However, there are immanent problems in his account. In particular, it has been shown that the passage from the minimal conception of experience to "I think" as a condition of experience is not convincing. Thus, in his detailed study on the structure of the Kantian argument, R. Howell denies that self-consciousness can be established as an analytic implication of the minimal conception of experience, that is, as necessary for any conceptual awareness of sensible data. Even if the step from the identity of "I think" to the necessary use of the categories were valid, according to Howell, the argument would not refute the Humean sceptic, who insists on the image of experience as a sequence of data without any reference to objects and the identity of "I think", placing these within the realm of philosophical illusions.32

The result that the explicit principles of transcendental deduction cannot be derived from a minimal conception of experience – as required by Strawson's reconstruction – can be seen as a confirmation of its "factual" status. However, this result also creates a problem concerning the Kantian critique of Hume. In Hume's conception of experience, the principles of transcendental deduction – in particular, the identity of "I think" – are not included; therefore, it seems necessary to derive them from a concept of experience that is also accepted by the sceptic and, thus, to give a reduction ad absurdum of his position. Henrich excludes this strategy. Instead, in his reconstruction, the deduction argument seems to merely appeal to an awareness of the respective principles as fundamental conditions of our knowledge, even though these are ignored by the Humean sceptic. This is, for Henrich, the function of reflection and its evidence. Thus, since its very beginning, Kant's argument presupposes a standpoint different from the Humean conception of experience – it presupposes the acceptance of conditions that are not accepted by the Humean sceptic and cannot be derived from a more fundamental conception of experience.

In comparison to the project of deduction in its analytic reconstruction, this means a weakening of its argumentative function. Instead of refuting the Humean position from its own premises, Henrich's deduction argument presupposes an opposed perspective, which cannot be established by arguments and is based on the evidence of a reflection that is pre-philosophical and inherent to our common discourse. Consequently, the function of the deduction argument can only consist of establishing objectivity, as founded on a priori principles, as a necessary moment of this perspective. Transcendental deduction only reveals the implications of the internal perspective to which it appeals.

With this implication of Henrich's interpretation, the profile of its philosophical project becomes clearer. However, it is also necessary at this moment to point out the limits of the analogy between "facts" in the juridical sense and in the sense of transcendental deduction. The jurist has to objectively demonstrate his claims about the relevant facts (which is done by means of documents and testimony). A transcendental deduction, however, can only appeal to a certain point of view – the non-Humean perspective of 'reflection' – to establish its legitimating facts. Thus, it is no surprise that the conception of a certain "point of view", which is fundamental to transcendental philosophy, became central in the methodological discussion after the publication of Critique of Pure Reason. In this discussion, the characterisation of the principles of transcendental philosophy as "facts" soon seemed questionable.33 Henrich's interpretation has, without doubt, brought to the centre of the discussion an aspect that is decisive to understanding the Kantian argument; however, the difference between the juridical context, the starting-point from which Kant introduced his notion of deduction, and the philosophical context has to be considered in more detail in order to gain a sufficiently clear comprehension of the sense in which the concept of a "fact" can elucidate the status of the principles of transcendental deduction. 



ADICKES, Erich. Kants Kritik der reinen Vernunft. Berlin: Mayer & Müller, 1889.

BECK, Jacob S., Erläuternder Auszug aus den critischen Schriften des Herrn Prof. Kant. Dritter Band, welcher den Standpunkt darstellt, aus welchem die critischen Philosophie zu beurteilen ist. Riga 1796.

ERDMANN, Benno. Kants Kriticismus in der 1. und 2. Auflage der Kritik der reinen Vernunft. Leipzig: Voss, 1878.

FICHTE, Johann G., Wissenschaftslehre nova methodo, organizado por E. Fuchs, Hamburg: Meiner 1982.

HENRICH, Dieter. "Die Beweisstruktur von Kants Transzendentaler Deduktion". In: PRAUSS, Gerold (ed.). Kant. Zur Deutung seiner Theorie von Erkennen und Handeln. Köln: Kiepenheuer und Witsch 1973, p. 90-104.

______________. Identität und Objektivität. Eine Untersuchung über Kants transzendentale Deduktion. Heidelberg: Winter, 1976.

______________. "Die Identität des Subjekts in der transzendentalen Deduktion", in: OBERER, Hariolf und SEEL, Gerhard (ed.). Kant: Analysen – Probleme – Kritik. Würzburg: Könighausen & Neumann, 1988, pp. 39-70.

______________. "Kant's Notion of the Deduction". In: FÖRSTER, Eckart. Kant's Transcendental Deduction. Stanford: Stanford University Press, 1989, p. 29-46.

_____________. Grundlegung aus dem Ich. Untersuchungen zur Vorgeschichte des Idealismus. Frankfurt a. M.: Suhkamp, 2004.

HOWELL, Thomas C., Kant's Transcendental Deduction. An Analysis of Main Themes in His Critical Philosophy, Dordrecht/Boston: Kluwer 1992

KANT, Immanuel. Kritik der reinen Vernunft (1. ed., 1781; 2. ed. 1787). In: Preussische Akademie der Wissenschaften (ed.). Kantswerke. Berlin: Walter de Gruyter, vol. III, 1904.

____________. Kritik der praktischen Vernunft (1. ed. 1788). In: Ibid., vol. V, 1908.

PATON, Herbert James. Kant's Metaphysic of Experience. London: Allen & Unwin, 1936, vol. I, p. 501.

STRAWSON, Peter F., The Bounds of Sense. London 1966.

VLEESCHAUWER, Herman de. "La déduction transcendentale dans l'oeuvre de Kant", in Werken uitgegeven door de Faculteit der Wijsbegeerte en Letteren. Antwerpen: De Sikkel, 1937. Vol. 3.



1 HENRICH, Dieter. Grundlegung aus dem Ich. Untersuchungen zur Vorgeschichte des Idealismus. Frankfurt a. M.: Suhkamp, 2004.
2 Ibid. "Kant's Notion of Deduction". In: FÖRSTER, Eckart. Kant's Transcendental Deduction. Stanford: Stanford University Press, 1989, p. 29-46.
3 Ibid. Kritik der praktischen Vernunft (1. ed. 1788). In: Ibid., vol. V, 1908.
4 KANT, Immanuel. Kritik der reinen Vernunft (1. ed., 1781; 2. ed. 1787). In: Preussische Akademie der Wissenschaften (ed.). Kantswerke. Berlin: Walter de Gruyter, vol. III, 1904.
5 Ibid. "Die Beweisstruktur von Kants Transzendentaler Deduktion". In: Prauss, Gerold (ed.), Kant. Zur Deutung seiner Theorie von Erkennen und Handeln. Köln: Kiepenheuer und Witsch 1973, p. 90-104.
6 Ibid. Identität und Objektivität. Eine Untersuchung über Kants transzendentale Deduktion. Heidelberg: Winter, 1976.
7 Ibid, op. cit. 1989, p. 29-46.
8 Ibid, op. cit., 1973.
9 Ibid, p. 91 ff.
10 ADICKES, Erich. Kants Kritik der reinen Vernunft. Berlin: Mayer & Müller, 1889.
11 PATON, Herbert James. Kant's Metaphysics of Experience. London: Allen & Unwin, 1936, vol. I, p. 501.
12 ERDMANN, Benno. Kants Kriticismus in der 1. und 2. Auflage der Kritik der reinen Vernunft. Leipzig: Voss, 1878.
13 VLEESCHAUWER, Herman de. "La déduction transcendentale dans l'oeuvre de Kant", in Werken uitgegeven door de Faculteit der Wijsbegeerte en Letteren. Antwerpen: De Sikkel, 1937. Vol. 3.
14 HENRICH, op. cit., 1973, p. 93.
15 Ibid, p. 94.
16 Ibid, p. 95 ff.
17 Ibid, op. cit., 1976.
18 The central argument of this essay is also formulated in Ibid,. "Die Identität des Subjekts in der transzendentalen Deduktion", in: OBERER, Hariolf und SEEL, Gerhard (ed.). Kant: Analysen - Probleme – Kritik. Würzburg: Könighausen & Neumann, 1988, pp. 39-70.
19 Ibid, p. 34 ff.
20 Ibid, p. 43 ff.
21 Ibid, p. 62 ff.
22 Ibid, p. 76 ff.
23 Ibid, p. 86.
24 Ibid, p. 94 ff. The argument sketched above does not yet establish specific rules in accordance with which the synthesis must be exercised. Regarding this point, Henrich observes in 1976 that deduction alone yields conditions for a system of rules without specifying what these rules would be; the system would demand reference to judgment forms as an independent premise (see p. 105). In 1988, however, Henrich expounds the three rules that correspond to the relation categories as implications of a system of a priori transition rules, thereby adding (1976) a justification of specific rules that does not need a reference to the judgment forms (see 1988, pp. 65-66).
25 Ibid, op. Cit., 1989.
26 Ibid., p. 30 ff..
27 Ibid., p. 34 ff..
28 The questions from which rises the transference of the concept of a "fact" to the transcendental context will be discussed in Section IV of this paper.
29 Ibid., p. 40 f..
30 See especially Strawson 1966, pp. 85 f.
31 See Strawson's contribution in Förster 1989, pp. 47 f..
32 See Howell 1992.
33 See the third volume of J.S. Beck's Erläuternder Auszug aus den critischen Schriften des Herrn Prof. Kant, which discusses the "point of view" presupposed in transcendental philosophy (Beck 1796). A conception of the appeal to the point of view of our epistemic activity, originally introduced by Beck, was adopted by Fichte in his second exposition of the principles of his system, the so-called "Wissenschaftslehre nova methodo" (Fichte 1982).